A method for constructing a mathematical model for automatic alignment of multiple optical paths and multiple optical targets

Abstract

The present invention provides a method for constructing a mathematical model of automatic light path alignment for multiple light paths and multiple optical targets in order to solve the problems that the existing mathematical model of light path alignment cannot be applied to multi-target and multi-light path automatic alignment, and cannot perform calibration of light path alignment parameters and statistics of alignment results in parallel. The present invention adopts a multi-dimensional vector operation mode in the derivation process of the mathematical model, and regards the parameter information of each light beam as a unit of the multi-dimensional vector, that is, the parameter information of 8 light beams is regarded as an 8-dimensional column vector, which participates in the calculation of all light path alignment target positions, pixel offsets, step deviations, collimation errors, and collimation state summary calculations, and constructs a mathematical model of automatic light path alignment for multiple targets, providing theoretical guidance for the realization of the light path docking alignment process.

Description

Technical Field

[0001] The invention relates to an optical path automatic alignment method, and in particular to an optical path automatic alignment mathematical model construction method for multiple optical paths and multiple optical targets. Background Art

[0002] The beam control and diagnosis system for large-scale devices is a multifunctional, high-precision laser parameter diagnosis platform, which includes a closed-loop control and monitoring system for optical sampling components, detection components, servo systems, monitoring systems, control systems, etc. It is used to accurately diagnose the characteristics of the laser beam output by the device, including energy, near field, far field, time waveform, etc. Under the condition that the back-end beam has been successfully collimated using the analog light source, the optical path docking collimation system mainly adjusts the front-end beam, and uses the main laser to track the analog light to adjust the main laser beam to the reference position where the analog light has been adjusted.

[0003] The large laser device contains a total of 8 light beams. Before each physical experiment, the 8 light beams need to pass through the optical path automatic alignment system to complete three alignment processes, including: optical path self-alignment, simulated light alignment and optical path docking alignment. According to the principle of optical path alignment proposed by Wang Zhengzhou et al. (Wang Zhengzhou, Wang Wei, Hu Bingliang, et al. Design and implementation of comprehensive diagnosis rapid automatic alignment system [J]. Acta Photonica Sinica, 2014, 43(5): 0512005-1~0512005-6 (EI: 20142617873991)) and Da Zhengshang et al. (Da Zhengshang, Li Dongjian, Zhou Wei, et al. Research on mathematical model of rapid adjustment of optical path automatic alignment [J]. Acta Photonica Sinica, 2008, 37(12): 2534-2538), the optical path alignment unit model mainly includes the following steps: 1) reading or identifying the optical target center of the alignment image; 2) converting the image deviation between the optical target and the reference into the number of X and Y direction motor steps required to adjust the optical target center to the target position; 3) moving the X and Y direction motors; 4) repeating steps 1) to 3) until the pixel deviation between the optical target center and the reference is less than the error threshold.

[0004] However, with the continuous deepening of physical experiments, the optical path docking and alignment process of large laser devices began to face some new problems:

[0005] 1) The number of optical targets has changed. The previous optical path docking and alignment process was to first turn off the simulated light source and then guide the main laser to achieve alignment. There was only one main laser target in the docking and alignment image. Now, in order to reduce the impact of the drift of the main laser target on the alignment accuracy, the simulated light source is no longer turned off in the optical path docking and alignment process, which makes one optical path docking and alignment image contain both the simulated light and the main laser optical targets.

[0006] 2) Changes in the mathematical model of optical path collimation. In the new mathematical model of optical path collimation, not only the characteristics of multiple optical targets should be reflected, but also the convergence conditions should be improved, that is, the convergence of the beam needs to determine whether the distance between the center of the main laser and the center of the target position is less than a given error threshold, rather than directly comparing the pixel deviation of the optical target center coordinates with the reference XY direction and the set error threshold.

[0007] 3) New requirements are put forward for alignment efficiency and time. Previously, the docking alignment process of large-scale devices was executed in series according to the optical path, which greatly affected the alignment efficiency. The parallel alignment method is the most direct and effective means to improve the optical path alignment efficiency and reduce the alignment time.

[0008] Specifically, for the optical path alignment mathematical model, in the original large-scale laser device, the automatic alignment of 8 beams all uses the same motor unit alignment model, which has obvious shortcomings:

[0009] 1) The original mathematical model has a unique optical target for the collimation image, which cannot meet the requirements of the collimation model when the collimation image contains multiple optical targets in the optical path docking collimation process;

[0010] 2) The original mathematical model is only for a single light beam, and the parameter information of multiple light beams is not reflected in the mathematical model, which is not conducive to the parallel calibration of multi-light path collimation parameters and the statistics of collimation results;

[0011] 3) The original mathematical model does not reflect the recognition algorithm for multiple optical targets, and cannot identify the simulated light target and the main laser target, thus failing to complete the collimation work.

[0012] Aiming at the shortcomings of the original collimation mathematical model, a new multi-light path and multi-target automatic collimation mathematical model is constructed to mainly solve the following problems:

[0013] 1) The new automatic alignment mathematical model can reflect the characteristics of multiple optical paths and multiple optical targets. In other words, under the premise of completing the multi-optical target recognition algorithm, the new multi-optical path and multi-target automatic alignment mathematical model needs to take the positions of the two optical targets, the main laser target and the simulated light target, as input conditions to meet the requirements of the alignment model for multiple optical targets;

[0014] 2) In the new collimation mathematical model, the derivation process of the mathematical model must adopt the operation method of multi-dimensional vectors. That is to say, in the mathematical model, the parameter information of each light beam is regarded as a unit of the multi-dimensional vector, and the parameter information of the 8 light beams is regarded as an 8-dimensional column vector, which participates in the operation of all light path collimation target positions, pixel offsets, step deviations, and collimation errors. Not only can the calibration of multi-light path collimation parameters be realized, but also after the collimation process is completed, the statistics of all light path collimation results can be realized.

[0015] 3) In the new collimation mathematical model, it is necessary to fully demonstrate the relationship between the multi-optical target recognition algorithm and the multi-optical path collimation mathematical model, so that the multi-optical target recognition algorithm becomes a necessary condition for the successful construction of the multi-optical path collimation mathematical model, providing a guarantee for the successful alignment of the optical path.

[0016] In order to solve the above problems, it is necessary to construct a new multi-light path and multi-target automatic alignment mathematical model. Summary of the invention

[0017] The purpose of the present invention is to solve the problem that the existing optical path collimation mathematical model is not applicable to multi-target and multi-optical path automatic collimation, and cannot perform the calibration of optical path collimation parameters and the statistics of collimation results in parallel, and to provide a method for constructing a mathematical model of optical path automatic collimation for multiple optical paths and multiple optical targets.

[0018] The basic design idea of ​​the present invention is to reflect the characteristics of multiple targets and multiple light paths in the mathematical model, that is, to adopt the multi-dimensional vector operation method in the derivation process of the mathematical model, and regard the parameter information of each light beam as a unit of the multi-dimensional vector, that is, to regard the parameter information of 8 light beams as an 8-dimensional column vector, which participates in the calculation of all light path collimation target positions, pixel offsets, step deviations, collimation errors, and collimation status summary calculations, and construct a multi-light path, multi-target automatic collimation mathematical model to provide theoretical guidance for the realization of the light path docking collimation process.

[0019] In order to achieve the above object, the technical solution adopted by the present invention is:

[0020] A method for constructing a mathematical model of automatic optical path alignment for multiple optical paths and multiple optical targets, which is special in that it includes the following steps:

[0021] Step 1), obtain k beams of light path simulation light collimation image f k (x, y) main laser target center and simulation light center; k ≥ 1;

[0022] (Base xy ,Center xy )=TargetRecognize(f k (x,y)

[0023] Among them, Base xy To simulate the center vector of the light target, Center xy is the main laser target center vector;

[0024] Step 2), calculate the deviation vector Δ between the main laser target center and the simulated light center xy :

[0025] Δ xy=Center xy -Base xy ;

[0026] Step 3) Calculate the collimation target position vector Target xy :

[0027] Target xy =Base xy +Offset xy

[0028] Among them, Offset xy To ensure the coaxial deviation vector between the main laser beam and the simulated beam, it is obtained through experimental calibration;

[0029] Step 4) Calculate the center of the main laser xy Move to the target location Target xy , the corresponding pixel deviation value vector ΔPixel on the CCD camera xy :

[0030] ΔPixel xy =Base xy +Offset xy -Center xy ;

[0031] Step 5) Calculate the center of the main laser xy Move to the target location Target xy , the motor step vector Step that needs to be adjusted xy :

[0032] ΔStep xy =ΔPixel xy ·*Ratio xy

[0033] Among them, Radio xy It is the proportionality coefficient between the number of motor steps in the X and Y directions and the number of pixels moved in the X and Y directions by the center of the main laser target on the CCD camera;

[0034] Step 6) Determine whether the single beam light path alignment is successful:

[0035]

[0036] Among them, AAResult k Indicates the alignment result. Success means the alignment is successful and the value is 1. Fail means the alignment is failed and the value is 0.

[0037] Step 7), repeating steps 1) to 6) until k light paths are successfully collimated, completing the automatic collimation of light paths of multiple light paths and multiple optical targets;

[0038]

[0039] Furthermore, in step 6), the single beam light path is successfully collimated if the following conditions are met:

[0040] (1) Center position of the main laser xy and target location Target xy Distance Less than the error threshold δ2;

[0041]

[0042] (2) The circular collimation of a single beam light path reaches convergence.

[0043] Furthermore, the error threshold δ2=5; the value range of the alignment cycle number l is [0-20).

[0044] Furthermore, in step 1), the k-beam light path simulation light collimation image f is obtained. k The method of determining the main laser target center and the simulated light center of (x, y) is a multi-target recognition method based on edge detection and least squares circle fitting.

[0045] Further, in step 1), 8≥k≥1.

[0046] Compared with the prior art, the present invention has the following beneficial technical effects:

[0047] The mathematical model of automatic optical path alignment for multiple optical paths and multiple optical targets provided by the present invention provides theoretical guidance for the implementation of the optical path alignment process, and can realize the calibration and calculation of various alignment parameters in the optical path alignment in parallel. At the same time, the present invention provides a basis for determining the success of the optical path alignment of 8 light beams by improving the convergence conditions, which has a significant guiding significance for the success of the optical path alignment process of large laser devices, and meets the requirements of precision and efficiency for the optical path alignment of large laser devices. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A flow chart of an embodiment of a method for constructing a mathematical model of automatic optical path alignment for multiple optical paths and multiple optical targets of the present invention;

[0049] Figure 2 A comparison diagram of multiple optical target recognition results during the optical path docking and collimation process of an embodiment of the present invention;

[0050] Figure 3The present invention provides an explanation of the operation process of the collimation mathematical model in the optical path docking and collimation process according to an embodiment of the present invention. DETAILED DESCRIPTION

[0051] In order to make the purpose, advantages and features of the present invention clearer, the following is a further detailed description of the method for constructing a mathematical model of automatic alignment of optical paths of multiple optical paths and multiple optical targets proposed by the present invention in conjunction with the accompanying drawings and specific embodiments. It should be understood by those skilled in the art that these implementations are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0052] The design idea of ​​the present invention is:

[0053] The mathematical model of multi-optical path and multi-optical target optical path collimation is the collimated image f k (x, y) is the input. On the premise of successful identification of the optical target, the center of the main laser target is xy Target xy The pixel deviation Δpixel xy Converted into the number of steps Δstep that the X and Y direction motors need to adjust xy , repeat the steps until the distance between the main laser center position and the target position is When the error is less than the error threshold δ2, and the main laser is guided to a fixed deviation position from the simulated light, the optical path docking and alignment process of the large laser device is completed.

[0054] like Figure 1 As shown, the method for constructing a mathematical model of automatic optical path alignment for multiple optical paths and multiple optical targets of the present invention specifically comprises the following steps:

[0055] Step 1) Obtain the collimated image of 8-way beam simulation light f k (x,y) center of main laser target and simulated light;

[0056] (Base xy ,Center xy )=TargetRecognize(f k (x,y)) (1)

[0057] Among them, Base xy To simulate the center vector of the light target, Center xy is the main laser target center vector;

[0058] Step 2), calculate the deviation vector Δ between the main laser target center and the simulated light center xy :

[0059] Δ xy =Center xy-Base xy (2)

[0060] Step 3) Calculate the collimation target position vector Target xy :

[0061] Target xy =Base xy +Offset xy (3)

[0062] Among them, Offset xy Offset is the coaxial deviation vector; xy It is a 1×2 dimensional vector, which indicates the deviations in the X and Y directions required to ensure the coaxiality between the main laser beam and the simulated beam, and can be obtained through experimental calibration;

[0063] Step 4) Calculate the center of the main laser xy Move to the target location Target xy , the corresponding pixel deviation value vector ΔPixel on the CCD camera xy :

[0064] ΔPixel xy =Base xy +Offset xy -Center xy (4)

[0065] During the alignment process of the optical path, the main adjustment is to adjust the center of the main laser. xy However, due to the drift of the simulated light beam, the simulated light target center Base xy It is not fixed, but changes over time in a small range. Therefore, the target position is a dynamically changing value, equal to the simulated light target center Base xy Offset xy sum.

[0066] Step 5) Calculate the center of the main laser xy Move to the target location Target xy The motor step vector Step that needs to be adjusted xy :

[0067] ΔStep xy =ΔPixel xy ·*Ratio xy (5)

[0068] Among them, Radio xyis the ratio coefficient between the number of motor steps in the X and Y directions and the number of pixels moved by the center of the main laser target on the CCD camera in the X and Y directions. xy It is a 1×2 dimensional vector with a floating point data type. When the proportional coefficient is greater than 0, it means that the motor moves in the same direction as the main laser target. When the coefficient is less than 0, it means that the two move in opposite directions.

[0069] Step 6) Determine the result of single beam light path alignment. Successful single beam light path alignment requires two conditions to be met:

[0070] (1) Determine the center position of the main laser xy and target location Target xy Distance Is it less than the error threshold δ2, that is δ2 is a given value;

[0071]

[0072] (2) The number of collimation cycles l for each optical path is less than 20.

[0073] The above single beam optical path collimation condition can be expressed by formula (7):

[0074]

[0075] Where l represents the number of collimation cycles, i.e., the order of beam convergence, and its value range is [0-20); AAResult k Indicates the collimation result. Success indicates successful collimation and its value is 1. Fail indicates failed collimation and its value is 0. k indicates the optical path number.

[0076] Step 7) Summarize the collimation results of the 8 beams.

[0077] For the optical path docking and alignment process of the entire device, each of the 8 optical paths needs to be aligned successfully to ensure the success of the alignment process of the entire device, that is, the optical path docking and alignment result of the entire device is the AND operation of the 8-path alignment results. Under the condition that the 8-beam optical paths are aligned successfully, k represents the optical path number:

[0078]

[0079] The following combination Figure 1 , the present invention is further described in detail through an embodiment.

[0080] Description of the process of constructing the mathematical model of multi-optical path and multi-optical target optical path alignment

[0081] Assume that in the simulation light collimation process, the image collected by each beam collimated CCD is f(x,y), (x,y) represents the image coordinates, the image size is 1600×1200, and the image format is 8-bit BMP image. The images collected by 8 beams of light path collimated CCD are used to construct a 1600×1200×8 multidimensional data cube, which can be expressed as:

[0082] f=(f 1 (x,y),f 2 (x,y),…,f 7 (x,y),f 8 (x,y) T (9)

[0083] The steps to construct a mathematical model for optical path alignment based on multiple optical paths and multiple optical targets are as follows:

[0084] Step 1), using a multi-target recognition method based on edge detection and least squares circle fitting, obtain the two target centers of the 8-way light beam simulated light collimation image;

[0085] Assume that the X-direction coordinate of the far-field center of the collimated image from optical path 1 to optical path 8 is The Y coordinate is The X-direction coordinates of the 8-way far-field centers are expressed as a one-dimensional column vector: The Y-direction coordinate is expressed as a one-dimensional column vector:

[0086] In order to facilitate matrix operations, the X and Y coordinates of the far-field center of the 8-way collimated image can be expressed using the two-dimensional column vector Base xy The corresponding matrix form is expressed as

[0087] Base xy =[B x ,B y ] (10)

[0088] Similarly, the X and Y coordinates of the 8-dimensional simulated light center are expressed using the two-dimensional column vector Center xy , and its corresponding matrix form is expressed as

[0089] Center xy =[C x ,C y ] (11)

[0090] In the formula, C x is a one-dimensional column vector, representing the X-direction coordinate values ​​of the centers of the eight simulated lights, C y It is a one-dimensional column vector, representing the Y-coordinate values ​​of the centers of the eight simulated lights.

[0091] Step 2) Calculate the deviation Δ between the far-field target center and the simulated light center of the 8 collimated images xy , the deviations of the far-field target center and the simulated light center in the X and Y directions of the 8 collimated images are expressed as:

[0092] Δ xy =Center xy -Base xy (12)

[0093] In the formula, Δ xy is a two-dimensional column vector, corresponding to an 8×2 matrix.

[0094] Step 3) Calculate the collimation target position Target xy

[0095] The purpose of simulated light collimation is to adjust the simulated light target center to the target position. The value of the target position is the far field center Base xy and the deviation value Offset xy The collimation target position is expressed by the formula:

[0096] Target xy =Base xy +Offset xy (13)

[0097] Where Offset xy It is a two-dimensional column vector, corresponding to an 8×2 matrix, and the first and second columns represent the target positions of the 8-way collimated images. xy Relative far field center Base xy Deviation values ​​in the X and Y directions.

[0098] Step 4) Calculate the simulated light center Center xy Move to the target center Target xy The pixel deviation that needs to be adjusted is expressed by the formula:

[0099] ΔPixel xy =Base xy +Offset xy -Center xy (14)

[0100] In the process of simulating the light path collimation, the collimated image of each light beam contains two targets, namely the simulated light center and the far field center. xy and the far field center Base xy They are all two-dimensional column vectors, corresponding to 8×2 matrices. Although the simulated light collimation mainly adjusts the simulated light centerxy , but the far field center Base xy It is not fixed and will change with the adjustment of the optical path. Therefore, the target position that needs to be adjusted for each path is a dynamically changing value, which is the far field center Base xy Offset xy sum.

[0101] Step 5) Calculate the simulated light center Center xy Move to the target center Target xy The number of motor steps to be moved ΔStep xy , which can be expressed as

[0102] ΔStep xy =ΔPixel xy ·*Ratio xy (15)

[0103] In the formula, Ratio xy Ratio is the ratio between the number of steps of the 8-way collimation motor in the XY direction and the center pixel deviation of the simulated light target. xy It is a two-dimensional column vector, corresponding to an 8×2 matrix, where the first and second columns are the ratios of the number of motor steps in the X and Y directions to the center pixel deviation of the simulated light target. xy The type of matrix elements is floating point number, and the positive and negative signs indicate the direction. In the X direction, "+" means that the number of steps of the X motor (positive number) is consistent with the X direction of the center of the simulated light target, and the target moves to the right; "-" means that the number of steps of the X motor (negative number) is opposite to the X direction of the center of the simulated light target, and the simulated light target moves to the left; in the Y direction, "+" means that the Y motor movement direction (positive number) is consistent with the Y direction of the center of the simulated light target, and the target moves upward; "-" means that the Y motor movement direction (negative number) is opposite to the Y direction of the center of the simulated light target, and the simulated light target moves downward.

[0104] Step 6) Determine whether the optical path alignment is successful.

[0105] The successful alignment of the 8-beam optical path requires the following three conditions to be met:

[0106] Condition 1: For each optical path, it is necessary to determine the center position of the simulated light. xy and target location Target xy Distance Is it less than the error threshold δ2 (constant value, the same for each optical path, the default value is 5), that is,

[0107] In the formula for:

[0108]

[0109] Condition 2: The number of collimation cycles l of each optical path is less than 20, that is, l < 20. For a beam convergence thread, moving the simulated light center to the target position requires multiple movements of the XY motor to achieve beam convergence.

[0110] The main reasons are: First, because the motor moves the number of steps ΔStep xy Center xy With Target xy The coefficient of deviation between the two is achieved through calibration, and there will be a certain error in the calibration coefficient. Secondly, because the mirror frame equipped with a stepper motor is a mechanical device, the motor will have a slight rigid oscillation during movement and stop. Compared with the micron-level oscillation of the motor, it appears as a pixel-level error in the CCD image, especially when ΔPixel xy When it is less than 3 pixels, it will oscillate back and forth around the target position, and multiple corrections are required to move the simulated light target to the target position.

[0111] In summary, the conditions for successful collimation of each optical path can be expressed as

[0112]

[0113] Where l represents the number of collimation cycles for each optical path, AARestlt k Indicates the collimation result of each optical path. Success indicates successful collimation, with a value of 1; Fail indicates failed collimation, with a value of 0.

[0114] Condition 3: For the simulated light collimation process of the entire device, each of the 8 light paths must be collimated successfully to determine that the collimation process of the entire device is successful, that is, the simulated light collimation result of the entire device is the AND operation of the 8-path collimation results. It can be expressed as:

[0115]

[0116] Experimental verification and result analysis

[0117] 1. Automatic alignment sequence of optical path

[0118] The control of the optical path automatic alignment process of large laser devices is completed through the optical path automatic alignment software, which completes the control of three alignment processes (optical path self-alignment, simulated light alignment, and optical path docking alignment).

[0119] For each collimation process, two control modes need to be completed: manual mode and automatic mode. In manual mode, you can select the start step and end step in the collimation process to execute a single or multiple steps to complete the collimation process of the optical path, which is convenient for adjusting and controlling the optical path. In automatic mode, you do not need to select the collimation process steps, and you can automatically execute each step of the collimation, which is convenient for automatic adjustment and control of the optical path.

[0120] When selecting the alignment order, the alignment start project must be less than or equal to the alignment end project, and the subsequent alignment projects can only be run after the previous alignment projects have been successfully run. The optical path alignment process consists of 8 steps, each of which has a strict order. The alignment order is as follows:

[0121] (1) Collimated simulation light is turned off;

[0122] (2) Automatically align all service tests;

[0123] (3) Turn on the tripled frequency analog light source laser;

[0124] (4) Opening the laser shutter of the triple frequency light source;

[0125] (5) Adjust the brightness of the triple frequency simulated light source (illumination);

[0126] (6) Move the motor to the collimation position [negative lens 2 out (MNG)];

[0127] (7) Collect the triple frequency CCD to find the center and adjust BM6 to the target position (move the motor...)

[0128] (8) Turn off the tripled frequency analog light source laser.

[0129] In manual operation mode, you need to open the simulated light collimation dialog box in the sub-controller interface and select the collimation order.

[0130] In the manual-automatic mode, select "Collect triple frequency CCD to find the center, adjust BM6 to the target position (move the motor...)" for the start of the collimation, and select "Collect triple frequency CCD to find the center, adjust BM6 to the target position (move the motor...)" for the end of the collimation, which means that the other steps have been successfully completed, and the core beam convergence process of "collecting images to find the center, moving the motor to the target position" is directly run. Click the "OK" button, and the optical path docking collimation begins.

[0131] 2. Relationship between multi-optical target recognition algorithm and multi-optical path collimation mathematical model

[0132] In order to illustrate the target recognition effect of the multi-optical target recognition algorithm in the optical path docking alignment process, this implementation compares the multi-optical target recognition results of 16 collimated images of 8 beams at the alignment start stage and the alignment success stage. Figure 2 As shown in the figure, at the beginning of the collimation, the 8-way beam collimation CCD collects images as shown in the first column, and the multi-target recognition results are shown in the second column. At the successful collimation stage, the 8-way beam collimation CCD collects images as shown in the third column, and the multi-target recognition results are shown in the fourth column.

[0133] Taking the optical path S1 as an example, the optical target recognition result at the beginning of the collimation is: the simulated light target corresponds to the reference Base xy (1,:), the coordinates are (507.24,245.86), marked with a red cross “+”, the main laser target corresponds to the center xy (1,:), the coordinates are (264.5,253.9), and are marked with a green cross “+”. After obtaining the simulated light target position and the main laser target position, the offset of the position adjusted according to the main laser target xy (1,:)=(122,57), the main laser target needs to be adjusted to the target position Target xy (1,:) coordinates are (629.24,302.86). The main laser target is adjusted to the target position Target xy (1,:) The pixel difference corresponding to the main laser target needs to be moved is Pixel xy (1,:)=(364.74,48.96), then the main laser target is adjusted to the target position Target xy (1,:) The number of BM6 motor steps to be moved xy (1,:)=(-91.18514.003).

[0134] For the 8-beam collimated light path, the multi-target recognition results of the 8 collimated images at the beginning of the collimation are shown in Table 1. The second and third columns are the corresponding reference Bases of the simulated light targets of the 8 collimated images. xy The center corresponding to the main laser target xy , the third column is the offset of the position adjusted by the main laser target xy , the 4th column is the main laser target and needs to be adjusted to the target position Target xy Column 5 is the pixel difference corresponding to the main laser target to be moved when the main laser target is adjusted to the target position. xy Column 6 is the BM6 motor step required to adjust the main laser target to the target position. xy For multi-optical target recognition algorithms, Base xyand Center xy is the return of the multi-optical target recognition algorithm of 8 collimated images; and for the multi-light path collimation mathematical model, Base xy and Center xy The input of the multi-path collimation mathematical model is Target xy is the intermediate calculation result of the multi-light path collimation mathematical model, Pixel xy Step xy It is the final output result of the multi-light path collimation mathematical model.

[0135] Table 1 Parameter calculation of the multi-optical target optical path docking collimation model (collimation start stage)

[0136]

[0137] Similarly, in the successful collimation stage, the optical target recognition result is: the simulated light target corresponds to the reference Base xy (1,:), the coordinates are (507.24,245.86), marked with a red cross “+”, the main laser target corresponds to the center xy (1,:), the coordinates are (633.1,299.5), and are marked with a green cross “+”. After obtaining the simulated light target position and the main laser target position, the offset of the main laser target position is adjusted. xy (1,:)=(122,57), the main laser target needs to be adjusted to the target position Target xy (1,:) coordinates are (629.24,302.86). The main laser target is adjusted to the target Target xy (1,:) The pixel difference corresponding to the main laser target needs to be moved is Pixel xy (1,:)=(-3.76,3.86), then the main laser target is adjusted to the target position Target xy (1,:) The number of BM6 motor steps to be moved xy (1,:)=(0.94,1.1040). It should be noted that in the successful alignment stage, since the main laser target is adjusted to the target position Target xy (1,:) The absolute value of the pixel difference corresponding to the moving main laser target is less than 4, where abs(-3.76)=3.76<4 in the X direction and abs(3.86)=3.86<4 in the Y direction, meeting the requirement that the collimation error is less than 5 pixels. This indicates that the beam convergence process has been completed and the optical path S1 docking and collimation has been successful.

[0138] For the 8-beam collimation light path, the multi-target recognition results of the 8 collimated images in the successful collimation stage are shown in Table 2. The second and third columns are the corresponding reference Bases of the simulated light targets of the 8 collimated images. xy The center corresponding to the main laser target xy , the third column is the offset of the position adjusted by the main laser target xy , the 4th column is the main laser target and needs to be adjusted to the target position Target xy Column 5 is the pixel difference corresponding to the main laser target to be moved when the main laser target is adjusted to the target position. xy Column 6 is the BM6 motor step required to adjust the main laser target to the target position. xy .

[0139] Table 2 Parameter calculation of multi-optical target optical path docking collimation model (collimation success stage)

[0140]

[0141] exist Figure 2 The bottom line shows the relationship between the multi-light path collimation mathematical model and the optical target recognition algorithm, that is, the simulated light target Base obtained by optical target recognition xy and main laser target Center xy It is the input parameter of the mathematical model of multi-optical path alignment. It is the basis for the number of motor steps in the XY direction of the BM6 two-dimensional mirror frame required for the convergence of the optical paths. It is an important guarantee for the success of the final optical path alignment.

[0142] At the beginning and successful stages of collimation, the eight light beams need to collect and identify the optical targets of 16 collimation images respectively. The multi-optical target recognition algorithm proposed in this embodiment is used to complete the optical path docking and collimation process of the eight light beams of the large laser device. In addition, since the convergence process of each light beam cannot be completed in one time, the beam convergence process of each light path is an average of 5 times, the average collection and multi-target recognition algorithm runs 5 times, and the eight light beams run the target recognition algorithm for a maximum of 40 times. This experiment completed the automatic collimation of the eight light beams by an average of 3 beam convergences, and the eight light beams ran the target recognition algorithm for a maximum of 24 times. This shows that the multi-optical target recognition algorithm proposed in this article is very effective in terms of the repeatability accuracy of target recognition.

[0143] pass Figure 2 From the analysis in , it can be seen that the relationship between the multi-light path collimation mathematical model and the optical target recognition algorithm is that the multi-optical target recognition algorithm provides input parameters for the multi-light path collimation mathematical model. In addition, the relationship between the two also has the following characteristics:

[0144] 1) The multi-optical path mathematical model provides an application scenario for the multi-optical target recognition algorithm;

[0145] 2) Both are to ensure the smooth completion of the optical path docking and alignment process;

[0146] 3) The matrix representation of the multi-path collimation mathematical model provides more reliable theoretical guidance for the automatic collimation of large laser devices with more light paths (48 beams).

[0147] 3. Explanation of the calculation process of the collimation mathematical model during the optical path docking and collimation process

[0148] In order to further illustrate the calculation process of the multi-path collimation mathematical model, show the changes of various parameters during this experiment, and further decompose the calculation process of the collimation mathematical model during the light path docking collimation process, the following figure is used to illustrate the calculation process of the multi-path collimation mathematical model. Figure 3 shown.

[0149] Figure 3 The change process of each parameter in the multi-path collimation mathematical model of 8-path beams in the collimation start stage and the collimation success stage is selected. The first line is the collimation start stage, the second line is the collimation success stage, and the third line is the mathematical model description; the red arrow “—>” marks the calculation process of each parameter in the multi-path collimation mathematical model, with a total of 10 columns, divided into 5 categories: input parameter vector, calibration parameter vector, output parameter vector, collimation result vector, and 8-path collimation result summary.

[0150] 1) Input parameter vector

[0151] There are two 8*2 dimensional column vectors, which are the corresponding bases of the simulated light target Base xy , the main laser target corresponds to the center xy ,exist Figure 3 The red frame mark is the output result of the optical target recognition algorithm and is also the input of the multi-light path collimation mathematical model. In other words, the output of the multi-optical target recognition algorithm marked with a red frame is the input of the multi-light path collimation mathematical model.

[0152] 2) Calibration parameter vector

[0153] There are two 8*2 dimensional column vectors, namely the coaxial deviation Offset xy , the ratio coefficient between the number of motor steps and the image pixels Ratio xy ,exist Figure 3 As shown in the black box. These two vectors are also auxiliary input parameters of the multi-path collimation mathematical model. Before the light paths are aligned, the two parameters of each light path need to be calibrated. The calibration accuracy is of great significance to the success of the light paths alignment.

[0154] 3) Output parameter vector

[0155] There are 3 8*2 dimensional column vectors in total, which are the main laser target that needs to be adjusted to the target position Target xy , the pixel difference corresponding to the moving main laser target Pixel xy , main laser target is adjusted to the target position Target xy Steps of BM6 motors to be moved xy .exist Figure 3 The blue box is used to mark the output of the multi-path collimation mathematical model. The calculation process of the three output parameter vectors is:

[0156] (1)Target xy =Base xy +Offset xy , corresponding to the collimation mathematical model formula (3);

[0157] (2)Pixel xy =Base xy +Offset xy -Center xy , corresponding to the collimation mathematical model formula (4);

[0158] (3)Step xy =Pixel xy *Ratio xy , corresponding to the collimation mathematical model formula (5);

[0159] It can be seen that Base xy and Center xy The output of the multi-optical target recognition algorithm is used as the input of the multi-light path mathematical model, which is marked with a red box; Offset xy and Ratio xy It is a parameter that needs to be calibrated in advance in the mathematical model, marked with a black box; Target xy 、Pixel xy and Step xy The calculation results obtained by outputting the multi-optical target recognition algorithm and pre-calibrating the parameters are marked with blue boxes.

[0160] At the beginning and successful stages of alignment, Base xy are the same, which means that the simulated light target remains unchanged. xy is different, indicating that the main laser target position has changed after the beam convergence process. xy and Target xy, the difference between the XY positions represented by the two vectors is large, Pixel xy and Step xy The pixel and motor deployment that needs to be adjusted is also huge.

[0161] 4) Collimation result vector

[0162] There are 2 8*1 dimensional column vectors, representing the distance between the laser center position and the target position. Single beam alignment result AAResult k , and the 8-beam light path collimation result summary AAResult all The calculation process of the two collimation result vectors is:

[0163] (1) Corresponding to the collimation mathematical model formula (6), k represents the optical path number, with a value range of [1 8], l represents the number of collimation cycles, with a value range of [0-20);

[0164] (2) Corresponding to the collimation mathematical model formula (7), since the maximum pixel deviation in the X and Y directions is equal to 4, the distance between the laser center position and the target position is error At the beginning of the collimation phase, the 8 beams All greater than 5.656, all AAResult k ==fail(1<=k<=8), in the collimation success stage, the 8-way beam All are less than 5.656, all AAResult k ==Sucess(1<=k<=8).

[0165] Compare the Center of the successful alignment stage xy and Targe txy The deviation between the XY positions represented by the two vectors is very small, and the pixel deviation that needs to be adjusted to move the main laser center to the target position is Pixel xy This means that after the beam convergence process, the main laser target center xy Already adjusted to the target position Target xy , the distance between the center position of each optical path laser and the target position The requirements of optical path docking and collimation for collimation accuracy have been met, which can be expressed as

[0166] 5) Summary of 8-way alignment results

[0167] The 8-way alignment result summary is a bool type value, Success indicates success (true), Fail indicates failure (false), and the calculation formula is:

[0168] The corresponding collimation mathematical model formula is (8).

[0169] At the beginning of alignment, for all AAResult k (1<=k<=8) performs AND operation, expressed as AAResult all ==false, the optical path alignment status is "Aligning", indicated by a yellow indicator light.

[0170] In the successful alignment stage, for all AAResult k (1<=k<=8) performs AND operation, expressed as AAResult all ==true, indicating that the optical path docking alignment process of the large laser alignment has been successfully completed, and the optical path docking alignment status is "successful", which is indicated by a green indicator light. Figure 3 The last column shows.

[0171] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the present invention.

Claims

1. A method for constructing a mathematical model of automatic optical path alignment for multiple optical paths and multiple optical targets, characterized in that: The following steps are involved: Step 1), obtain k beams of light path simulation light collimation image f k (x, y) main laser target center and simulation light center; k ≥ 1; (Base xy ,Center xy )=TargetRecognize(f k (x,y)) Among them, Base xy To simulate the center vector of the light target, Center xy is the main laser target center vector; Step 2), calculate the deviation vector Δ between the main laser target center and the simulated light center xy : Δ xy =Center xy -Base xy ; Step 3) Calculate the collimation target position vector Target xy : Target xy =Base xy +Offset xy Among them, Offset xy To ensure the coaxial deviation vector between the main laser beam and the simulated beam, it is obtained through experimental calibration; Step 4) Calculate the center of the main laser xy Move to the target location Target xy , the corresponding pixel deviation value vector ΔPixel on the CCD camera xy : ΔPixel xy =Base xy +Offset xy -Center xy ; Step 5) Calculate the center of the main laser xy Move to the target location Target xy , the motor step vector Step that needs to be adjusted xy : ΔStep xy =ΔPixel xy ·*Ratio xy Among them, Radio xy It is the proportionality coefficient between the number of motor steps in the X and Y directions and the number of pixels moved in the X and Y directions by the center of the main laser target on the CCD camera; Step 6) Determine whether the single beam light path alignment is successful: Among them, AAResult k Indicates the alignment result. Success means the alignment is successful and the value is 1. Fail means the alignment is failed and the value is 0. Step 7), repeating steps 1) to 6) until k light paths are successfully collimated, completing the automatic collimation of light paths of multiple light paths and multiple optical targets; 2. The method for constructing a mathematical model of automatic optical path alignment for multiple optical paths and multiple optical targets according to claim 1, characterized in that: In step 6), the single beam optical path is successfully collimated if the following conditions are met: (1) Center position of the main laser xy and target location Target xy Distance Less than the error threshold δ2; (2) The circular collimation of a single beam light path reaches convergence.

3. The method for constructing a mathematical model of automatic optical path alignment for multiple optical paths and multiple optical targets according to claim 2, characterized in that: The error threshold δ2=5; the value range of the alignment cycle number l is [0-20).

4. The method for constructing a mathematical model of automatic optical path alignment for multiple optical paths and multiple optical targets according to claim 1 or 2, characterized in that: In step 1), the k-beam light path simulation light collimation image f is obtained. k The method of determining the main laser target center and the simulated light center of (x, y) is a multi-target recognition method based on edge detection and least squares circle fitting.

5. The method for constructing a mathematical model of automatic optical path alignment for multiple optical paths and multiple optical targets according to claim 1, characterized in that: In step 1), 8≥k≥1.

Images